Intel CFO David Zinsner detailed a strategic "course correction" focused on demand-based investment and increased transparency with partners to improve yields. The company highlighted strong progress on the 18A process node, with Panther Lake yields tracking ahead of schedule. While server CPU demand has rebounded significantly—driven by AI orchestration needs—Intel faces severe supply constraints across wafers and substrates, with Q1 marking the trough for shipment shortages. Management reaffirmed its target to exit 2027 with breakeven operating margins for Foundry, bolstered by a multi-billion dollar pipeline in advanced packaging opportunities.

Key Takeaways

• 18A Progress: The 18A process is performing at or "slightly better" than the yield glide path, with Panther Lake demand currently outstripping supply.
• Server Market Recovery: Server unit TAM grew "mid-20s-percent" last year and is expected to increase "meaningfully" again this year as CPUs regain relevance for AI agentic workloads.
• Supply Constraints: Q1 represents the worst period for under-shipping demand; supply constraints for wafers, substrates, and memory are expected to persist "certainly this year."
• Foundry Breakeven: Intel maintains its guidance to "exit 2027 at breakeven operating margins for foundry," supported by operational efficiencies and external wins.
• Packaging Revenue Upside: Expectations for advanced packaging (EMIB-T) wins have shifted from the "hundreds of millions" to closing deals in the "billions of dollars per year."
• Gross Margin Targets: The near-term objective is to improve gross margins through the year until they "start with a 4%" (40% range).
• Capital Expenditures: CapEx for the year is guided "flattish," with increased tool spending offsetting a decrease in shell/space spending.
• Memory Headwinds: Memory shortages are expected to last through next year, potentially constraining client PC growth in the second half of the year.

Q&A

Joe Moore (Morgan Stanley): Can you describe the big picture regarding the strategy since the leadership transition and whether these are major course corrections?

• David A. Zinsner: It is primarily a course correction where investments are now tied to "real signals" of demand rather than getting ahead of it. Leadership has simplified the organization by reducing layers and, crucially, has "opened up a lot of data to our partners," which has been a key factor in improving 18A yields.

Joe Moore (Morgan Stanley): How is the 18A process progressing, particularly with Panther Lake as a proof point?

• David A. Zinsner: The process is at or slightly better than its yield glide path, and progress is "probably a bit ahead of schedule." While originally viewed as an internal node, the success of 18A has generated inbound interest to offer it as a foundry node (18A-P) to external customers.

Joe Moore (Morgan Stanley): How durable is the incremental demand you are seeing for server CPUs in the context of AI?

• David A. Zinsner: CPUs have "become cool again" as workloads shift from running LLMs to orchestration and agentic tasks which run on CPUs. Server unit TAM was up "mid-20s-percent" last year, is expected to rise meaningfully this year, and customers are now seeking long-term agreements of three to five years.

Joe Moore (Morgan Stanley): Why are supply constraints so severe right now, and is Q1 the worst point for under-shipping relative to demand?